Fast fourier transform based new pooling layer for deep learning

Convolution is considered most significant layer in deep learning because it can extract best features of data through the network but it may result in huge volume of data. This problem can be solved by using pooling. In this paper, A novel pooling method is proposed by using discrete Fourier transform (DFT), this method is used DFT technique to transform the data from spatial domain into frequency domain to preserve the most important information from the details coefficients, where the details information of the image is less significant, therefore it can be discarded to down sample the size of dimensions. Its effect will be great with advantage of reducing the eliminated details information as compared with other standard methods. After applying DFT, the most significant coefficients, which represent most important features are cropped while less important details will be discarded then the data are reconstructed by applying inverse DF, therefore the high quality of features are extracted, which solve the problem of losing significant information during the pooling layer. Different methods are proposed based on the scenario of using DFT. The proposed methods are tested by extracting pooled image then the original images were retrieved using only the pooled images. Then the retrieved images are compared with original images by using different measures such as SNR, correlation and SSIM. Then the proposed layers used for image classification for two different datasets. The results proved that the proposed methods outperformed standard methods, thus it can be used for deep learning application.